Last time my back bit me I used rolling stool I usually use when I am abusing myself working on the motorcycle. In any case I was able to lower it so my work was close to eye level on bench. It seemed to alleviate pain don't know why I don't do that all the time. In any case take pain pills (don't feel guilty if you enjoy the buzz) and if you are in the mood to hallucinate- hallucinate a servo for the pumkin. Just kidding get well soon.
Bill
Bill
GuidoR said:Hi Choky!
Still waiting for the "Cook Book"
After a lot of thinking on LDR matching I now need some "real life" experimentation...
thanks!!!
back in a saddle in last few days
I'm trying to solve several most urgent things , including packing some still not delivered PSM LS-es .
Cook Book is in final stage , but I must sit by my PC and retype it , which isn't an option for more than 15mins at least one week more .
sitting is baaad for back
greetings from spineless ZM
Zen, I'm also a lot better now & ready for some diy (will I never learn).
I have some parts shipping to do a preamp using unmatched LDRs & microprocessor. If successful, this could open the way to mutichannel or balanced operation.
I thought of using the uC to run a calibration function in which all the LDRs would be automatically measured at various currents and the offset from an ideal would then be stored in an array for use during operation (see the graph attached which came from jackiinj to see how an LDR differs from it's ideal)
After a bit of discussion about temp drift,etc., it occurred to me that maybe dynamic matching of the LDRs while in operation would be more accurate. I envisage this dynamic matching taking place for short time after a volume change but not continuously.
What is the best way to sense the resistance of the LDR during operation, without it affecting the sound?
I have some parts shipping to do a preamp using unmatched LDRs & microprocessor. If successful, this could open the way to mutichannel or balanced operation.
I thought of using the uC to run a calibration function in which all the LDRs would be automatically measured at various currents and the offset from an ideal would then be stored in an array for use during operation (see the graph attached which came from jackiinj to see how an LDR differs from it's ideal)
After a bit of discussion about temp drift,etc., it occurred to me that maybe dynamic matching of the LDRs while in operation would be more accurate. I envisage this dynamic matching taking place for short time after a volume change but not continuously.
What is the best way to sense the resistance of the LDR during operation, without it affecting the sound?
jkeny said:
There is probably a better way, but what about having a "Power on" calibration, followed by the option to push a button (could be an encoder with push switch), this could power a relay that switches out the incoming audio, and replaces it with whatever you wish to use for calibration. Another option that comes to mind would be to monitor through a buffer, and when a silence of more than ? has passed it could run the calibration, if its fast enogth it could be done between tracks, most of us need time to chose the next album. To make flexible some options could be selected from the encoder push switch, for example with the pre amp controller i made myself I hold in the volume and turn to select the input, and if i push the volume knob in on power up I can change the step resolution (rotate the vol knob through the options and let it go when I have selected the option I wish).
Just ideas!
Good ideas, Stones - I do indeed need an initial full calibration and then some subsequent point or full calibration options as you pointed out - it will all depend on how quickly the calibration can be performed.
I don't need to use any signal to calibrate just a run through the vol settings(ie current sweep in 64 steps from 0 to 20mA ) & a measure of the resistance produced at each step on each LDR.
I don't need to use any signal to calibrate just a run through the vol settings(ie current sweep in 64 steps from 0 to 20mA ) & a measure of the resistance produced at each step on each LDR.
well - there is one well preserved secret
all you need is to measure each LDR in most common used currents - few points bellow 1mA and few points above that ( up to 10mA) ;
then sort half of them looking at low current/high resistance , and half of them looking at higher current/lower resistance .
use first for series ones , second ones for shunts
think about their functioning ......... that's good enough
we made several balanced and non balanced PSM LSes , with LDRs from first batch , which were matched pretty rough , comparing to second batch .......... and result are more than satisfactory .
anyway - whatever you do , common pots aren't too good balance wise .... so LS can't be worse so easy
all you need is to measure each LDR in most common used currents - few points bellow 1mA and few points above that ( up to 10mA) ;
then sort half of them looking at low current/high resistance , and half of them looking at higher current/lower resistance .
use first for series ones , second ones for shunts
think about their functioning ......... that's good enough
we made several balanced and non balanced PSM LSes , with LDRs from first batch , which were matched pretty rough , comparing to second batch .......... and result are more than satisfactory .
anyway - whatever you do , common pots aren't too good balance wise .... so LS can't be worse so easy
Could be a stupid question
But would there be any benefit from using double mono pots ?
Could be linked together behind the front
Some have done so simply by a rubber band that was easily adjusted
Anyway, I think I remember to have seen a volume attenuator which was like a normal single stereo pot, but made of two halves that could be adjusted seperately, but both moved together in normal use
Does this sound like rubbish ?
But would there be any benefit from using double mono pots ?
Could be linked together behind the front
Some have done so simply by a rubber band that was easily adjusted
Anyway, I think I remember to have seen a volume attenuator which was like a normal single stereo pot, but made of two halves that could be adjusted seperately, but both moved together in normal use
Does this sound like rubbish ?
Good info Zen, but just to make sure I've got it right, in this graph : http://picasaweb.google.com/udailey/LightspeedAttenuator#5292315840126431810 of 24 NSL-32SR2 tested with a 3V battery and the following resistors in series with the + side of the battery:
100R
9960R
22000R
47000R
do you mean to use the bunch at the top half as series LDRs & the bunch at the bottom half as shunt LDRs?
100R
9960R
22000R
47000R
do you mean to use the bunch at the top half as series LDRs & the bunch at the bottom half as shunt LDRs?
Tinitus, Not such a siily question at all - it works okay, you have a simple balance control included, just like the Dual Concentric Vol Pot on preamps years ago with the extra shaft in the middle to turn the rear pot. Only problem was, the 2 seperate elements are no better matched than any other dual pot pair and stereo balance isn't particularly good. Surprising thing is, despite what was obviously poor tracking between pots, it just didn't seem to matter that much and anything really noticeable was simply adjusted right. With the better results with the Lightspeed system, it is more critical. However, all is not lost as the method of matching pots as seen on the "esp" website (adding parallel resistor) will pull them together - maybe not ideal curve and all that theory stuff, but works like a treat. With georges original scheme, the 2 pots together was to avoid the one pot suffering deterioration due to too much thru it - the 2 sides do track together even tho' the perfect log attenuation curve maybe a bit suspect with the cheap pot jkeny, I think what ZM means is that most of us use a vol control in the lower position (about 9 - 10 'clock) that means a fair bit of attenuation and hence the series resistance is higher here and the shunt is lower. Thus, you can use 2 pairs of matched LDRs that are quite well apart on the chart of his - as you say, the top half and bottom half of the curves. Curiously enough, of the 2 dozen, there appers to be only a few that cut across other curves - the light green one and perhaps a khaki one - (rejects for another grouping). Most of the others will match perfectly well with the trimming pot. One thing with this setup, is to somehow add something to keep all the LDRs at same temp (copper strip and potting glue, etc) It'll work okay without, but why not get it right as the performance of this little bunch of "home brew" bits will totally outperform almost anything that you can buy retail on the market, regardless of cost, despite the inclusion of the debilating input switched, etc!.
Yes jameshill,
The curves mostly are the same slope but there is divergence between them - this divergence increases with increased attenuation - the top group ranges from 5000 to 12000 ohm - the bottom group ranges 2000 to 4000 ohm. How severe is this mismatch? I've no handle on this so just looking for guidance.
Even if adjusted with a 1K pot for achieving a balance between channels - is this balance only at one setting & at other settings there is still this mis-match
The curves mostly are the same slope but there is divergence between them - this divergence increases with increased attenuation - the top group ranges from 5000 to 12000 ohm - the bottom group ranges 2000 to 4000 ohm. How severe is this mismatch? I've no handle on this so just looking for guidance.
Even if adjusted with a 1K pot for achieving a balance between channels - is this balance only at one setting & at other settings there is still this mis-match
jkeny said:
hm .....
more like matching by left vs. right - looking from vertical situated at middle of that graph ;
if you sort all results ( say - for 4 different currents ) , you can see that around half of them are well sorted ( like a nice lineup ) for high resistance , and other half is pretty well sorted ( lineup) in low resistance values
so - take first ones as pairs or quads for series elements , and second ones as pairs or quads for shunt elements
try to fiddle with possible values in resistance divider ( imaging that you are turning vol knob ) and you'll see that that way you'll have pretty good tracking , just counting on fact that you have nearest neighbors for all series , same as for all shunt elements .
my ears are certainly more apart ......
pchw said:Do we see the SpeedLight in the tunnel yet
I certainly saw that light
I just finished some scans these days ( blah blah MRI) , and I'm happy camper - situation is not worse than on previous one ( 6 years ago) .......
anyway - last packages will be sent day after tomorrow ( Friday)
huh - dunno what was worse - my back , or feel because of delay in sending funny envelopes
